easysnowdata.processing.snow.snodas_members# easysnowdata.processing.snow.snodas_members(tar_bytes: bytes | BufferedIOBase) → dict[str, dict[str, bytes]][source]# Group a SNODAS day’s tar into {stem: {"dat": ..., "txt": ...}}. The members are gzipped inside the tar; both are decompressed here so the caller only deals with bytes.